Work
Odinzen brings together software engineering and computational materials science, doing original research in the field and building the scientific software, agents, and data systems around it. The work falls into three areas.
Software & AI#
Scientific software and agentic systems we’re building on open-source foundations. The agents call real solvers and check their own numbers against the science.
- Agentic automation of materials workflows — computational thermodynamics in the loop: CALPHAD assessments and phase-equilibrium analysis, with the tools, grounding, and guardrails scientific work needs.
- Custom scientific software — tooling and agentic systems we’re building on open-source libraries, contributed back where we can.
Data Systems#
We’re building systems and tools that gather and organize research data, so it can feed both people and agents.
- Knowledge graph — a structured representation of materials data suitable for machine reasoning.
- Data pipelines — turning scattered research data into something queryable and reproducible.
Strategy & Design#
We’ve shipped production software and done the lab science, and we share what we’re learning about bringing agentic methods into technical work without losing rigor or traceability.
- Speaking — talks and conference sessions on building agentic AI for scientific work: what we’re learning as we build it, how to keep rigor and traceability when the agents do the work, and where they still need us.
- Training — hands-on workshops for teams and research groups building their own agentic systems, drawn from the patterns we use in our own research.
- Engineering design — translating thermodynamic assessments into maintainable software and computational workflows.
How we work#
We work in the open. Odinzen’s roots are in open-source scientific software, and we’re building agentic systems on open foundations, contributing back to the ecosystem where we can.
Contact#
Reach out at info@odinzen.io or via LinkedIn to discuss a collaboration.